Feature Extraction Evaluation of Various Machine Learning Methods for Finger Movement Classification using Double Myo Armband

The deployment of electromyography (EMG) signals can be used in decoding finger movements for exoskeleton robotics, prosthetic hands, and powered wheelchairs and thus has attracted the attention of many researchers. However, decoding any movement is a challenging task. The success of using EMG signals depends on the appropriate choice of feature extraction and classification model, especially in the feature extraction process. Therefore, this study conducted an eight -feature extraction evaluation on various machine learning methods, i.e., Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), Decision Tree (DT), Naive Bayes (NB), and Quadratic Discriminant Analysis (QDA). Datasets from four intact subjects were used to classify twelve finger movements. Through five cross-validations, the result showed that almost all feature extractions combined with SVM outperformed the other combinations of features and classifiers. Mean absolute value (MAV) as a feature and SVM as a classifier were the best combination, with an accuracy of 94.01%.